62 results about "Expectation–maximization algorithm" patented technology

The invention discloses a multi-dimensional distribution offset adaptive threshold detection method and system, and the method comprises the steps: obtaining real-time data, extracting a multi-dimensional statistical feature, and obtaining a feature vector; based on historical normal data, using an expectation maximization algorithm to train a Gaussian mixture model, and determining parameters to obtain a normal distribution model; inputting the feature vector into the model, and calculating a probability value of the feature vector belonging to normal distribution as a first offset judgment index; based on the real-time data distribution of a plurality of detection objects in the same group, the distribution difference of any two objects is calculated by using a Wasserstein distance, and the similarity between the objects is obtained; and constructing a similarity network and calculating connectivity as a second offset judgment index. Setting a fixed-length sliding window, dynamically updating two indexes in the window, and obtaining a first self-adaptive threshold value and a second self-adaptive threshold value; and when any index is lower than a corresponding threshold value, determining distribution offset and giving an alarm, and updating model parameters in real time by using an incremental expectation maximization algorithm. According to the invention, accurate detection and intelligent analysis of data distribution offset are realized.

The invention discloses an Internet of Things anomaly detection method and system based on quaternion state space diffusion enhancement, and belongs to the technical field of network security and artificial intelligence. The method comprises the following steps: mapping a flow time sequence feature into a quaternion tensor to maintain an internal coupling relationship of a multi-dimensional feature; a double-flow encoder is designed, a quaternion selective state space model is adopted to extract continuous fluid features, and a dynamic hypergraph neural network is adopted to model discrete protocol features; carrying out self-supervised pre-training on a resistance pseudo-anomaly sample by utilizing potential diffusion model generation, and optimizing characterization by combining quaternion cepstrum distance loss; the injected learnable prompt vector is optimized in the small sample fine tuning stage, and a category prototype is corrected by using a semi-supervised expectation maximization algorithm; and calculating a sample anomaly score based on an energy model to realize known attack classification and unknown anomaly judgment. According to the method, the generalization ability of the model under the small sample condition and the unknown threat detection ability are remarkably improved.

The invention discloses a hybrid antenna array direction of arrival estimation method based on noise marginalization SBL. The method comprises the following steps: initializing system parameters and a sampling grid set; establishing a hybrid antenna array receiving signal model, and initializing a hybrid beam forming matrix into a block diagonal matrix meeting constant modulus constraint; constructing a hierarchical probability model under a Bayesian framework, and introducing a noise precision parameter into signal prior; integral operation is carried out on the noise precision parameter, so that the signal posterior distribution is converted into student t distribution; an objective function is constructed, an expectation maximization algorithm is adopted to iteratively update a signal energy spectrum, and the process does not involve noise parameter estimation; an alternating iteration strategy is adopted, the signal energy is fixedly updated through inner circulation, and optimization is performed through a gradient descent method after outer circulation is fixed; after convergence, constructing a target function of off-grid estimation by reconstructing a covariance matrix; and searching the off-grid direction of the maximized objective function near a spectrum peak to obtain a final DOA estimation value.

The invention discloses a frequency domain adaptive clutter suppression method and system for a dual-polarization weather radar, and relates to the technical field of weather radar processing. According to the invention, pulse-level quality control and ground feature identification marking are carried out on radar original echoes; calculating power spectrums of the marking units, performing dynamic noise measurement and data quality grading, and screening out effective processing units; constructing a Gaussian mixture model of a power spectrum of an effective unit, and iteratively estimating model parameters through an expectation maximization algorithm to realize complete separation of a ground feature spectrum and a meteorological spectrum in a frequency domain; inverting single-channel parameters based on the separated meteorological spectrum; a cooperative spectrum separation strategy is adopted for horizontal and vertical polarization channels, and dual-polarization parameters such as differential reflectivity, correlation coefficients and differential phases are inverted based on the filtered dual-channel signals; according to the invention, high-precision separation of clutters and meteorological echoes is realized, the loss of meteorological signals is significantly reduced while ground features are effectively suppressed, and the accuracy and reliability of dual-polarization parameter inversion are improved.

The invention discloses an intelligent flood forecasting method for coupling error correction and joint modeling, and relates to a deep learning and uncertainty modeling technology. At the input end, constructing a future random rainfall scene through hourly dynamic normal disturbance; the method comprises the following steps: at a model end, introducing a multi-structure and multi-objective function combination based on Kolmogorov-Arnold Networks and Transform, and forming a multi-member ensemble forecast; at an error end, a probabilistic modeling method based on a numerable asymmetric Laplacian mixed density network is provided, and fine error correction is realized; a Vine copula function is adopted to construct high-dimensional joint distribution, and a Bayesian model averaging and expectation maximization algorithm is combined to realize weighted fusion of multi-member posterior results; according to the method, uncertainty in flood forecasting can be comprehensively described, the stability and adaptability of a forecasting system are improved, and the method is suitable for a basin-level real-time flood ensemble forecasting scene.

The invention discloses a radar interference effect evaluation method based on a constraint learning dynamic Bayesian network, is applied to the field of radar interference evaluation, and aims at solving the problem that the accuracy of interference effect evaluation is reduced due to radar detection data missing in a complex electromagnetic environment. Meanwhile, parameter constraints of five types of evaluation indexes and interference effect grades are defined; secondly, constructing a constraint learning dynamic Bayesian network, and learning a conditional probability and a transition probability under a data missing condition; then, proposing a prior constraint expectation maximization algorithm, converting parameter learning into an optimization problem with constraint by combining convex optimization, and overcoming the defects of a traditional expectation maximization algorithm; secondly, a cloud model is introduced to quantify discrete probability distribution into a continuous interference degree value; finally, simulation shows that the method can effectively improve parameter learning stability and evaluation accuracy under the conditions of suppressing and deception jamming and single index deficiency, and provides a reliable scheme for radar jamming effect evaluation in a complex environment.

The invention discloses a multi-dimensional night urination behavior monitoring system based on rhythm features, and relates to the technical field of biomedical signal processing, and the system specifically comprises a data acquisition module, a rhythm feature extraction module, a urine volume dynamics modeling module, a clustering monitoring module and a real-time feedback module; collecting individual data in real time through IoT equipment and preprocessing the individual data; calculating a nocturnal urination frequency index, a nocturnal urination time concentration ratio, a urination interval rhythm variation coefficient and a deviation index based on the individual data, and performing rhythm feature extraction; performing quadratic polynomial least square fitting on the night accumulated urine volume, and calculating a urine volume acceleration index by using an obtained second derivative to quantify a urine volume generation trend; the method comprises the following steps: constructing and preprocessing a night urine multi-dimensional digital phenotypic vector matrix, fitting a Gaussian mixture model based on an expectation maximization algorithm of a Bayesian information criterion, automatically mapping individuals into four types of subtypes according to cluster center features, and outputting individual subtype labels and confidence coefficients; and obtaining a comprehensive risk score through normalized risk assessment.

The invention discloses a two-stage adaptive numerical distribution reconstruction method based on local differential privacy. And the terminal equipment applies a local differential privacy perturbation mechanism to the held numerical private data to generate perturbation data and sends the perturbation data to the aggregation server. After an aggregation server collects data, first-stage estimation is executed firstly, and initial smooth distribution with low variance characteristics is generated based on an EMS algorithm to serve as a guide map; then, a non-uniform self-adaptive bucket dividing strategy is constructed according to cumulative distribution characteristics of initial distribution, and equal probability interval division is achieved; and finally, second-stage estimation is executed, re-statistics is carried out on noise data based on a self-adaptive bucket dividing strategy, accurate reconstruction is carried out by applying an expectation maximization algorithm with space kernel smoothing, and weighted updating is carried out on threshold-controlled space smoothing by utilizing a kernel function based on a bucket center physical distance. According to the method, the problem of over-fitting of discretization deviation of uniform bucket division and an expectation maximization algorithm in a local differential privacy high-noise environment is effectively solved through a two-stage strategy, and the accuracy and robustness of value distribution estimation are remarkably improved.

The invention relates to the technical field of power systems, in particular to a wind-light and load typical scene optimization method. The method comprises the following steps: establishing a mathematical model of core equipment in the multi-energy coupling system; on the basis of the mathematical model, an improved Gaussian mixture model and an improved expectation maximization algorithm are adopted, and a typical scene set with probability distribution fitting reality is generated in combination with space-time correlation characteristics of wind and light output and loads; the method comprises the following steps of: constructing an optimal scheduling model containing multi-energy supply and demand balance constraints by taking system total cost minimization as a target and combining an operation boundary condition and a typical scene set of core equipment, and solving the optimal scheduling model after performing linearization processing on the optimal scheduling model to obtain a flexibility adjustment scheme of the multi-energy coupling system. According to the method, the influence of multiple uncertain factors on the system regulation capability can be comprehensively reflected, the modeling precision of the probability distribution of uncertain variables such as wind, light, load and the like is improved, and the accuracy of the flexibility evaluation of the power system and the rationality of resource allocation are improved.

The present application relates to a method for optimizing energy consumption of an electric-thermal integrated energy system with reduced loss, comprising: obtaining energy consumption data of the electric-thermal integrated energy system; processing the obtained energy consumption data, classifying it according to energy consumption regions, and dividing time periods; analyzing the energy consumption data of different regions respectively, obtaining consumption sources; optimizing the energy consumption of different regions respectively, generating improvement suggestions for the consumption sources; implementing loss reduction measures, and analyzing and comparing the loss reduction effects, recording the effective loss reduction measures for popularization and use. The present application also discloses an energy consumption optimization system. The present application obtains energy consumption data of different regions and different time intervals, monitors energy consumption abnormal nodes in real time; uses a normal distribution probability model of an expectation maximization algorithm to calculate the optimal energy consumption data distribution, generates improvement suggestions for the consumption sources, and optimizes energy consumption.

The application discloses an Internet of Things anomaly detection method and system based on quaternion state space diffusion enhancement, and belongs to the technical field of network security and artificial intelligence. The method comprises the following steps: mapping traffic time sequence features into a quaternion tensor to maintain the internal coupling relationship of multi-dimensional features; designing a double-flow encoder, extracting continuous flow features by using a quaternion selective state space model, and modeling discrete protocol features by using a dynamic hypergraph neural network; generating an adversarial pseudo-anomaly sample by using a latent diffusion model to perform self-supervised pre-training, and combining a quaternion cepstrum distance loss to optimize the representation; optimizing the injected learnable prompt vector in the small sample fine-tuning stage, and correcting the class prototype by using a semi-supervised expectation maximization algorithm; calculating sample anomaly scores based on an energy model to realize known attack classification and unknown anomaly determination. The application significantly improves the generalization ability of the model under the condition of small samples and the detection ability of unknown threats.

The invention belongs to the technical field of information and communication, and relates to an active user, time delay and channel joint estimation method for a low-precision receiver. The problem that when a receiver adopts a one-bit analog-to-digital converter, transmission delay and channel state estimation precision are too low due to high quantization errors is solved. Performing correlation peak detection on the received signal and the known leader sequence, and preliminarily estimating active users and transmission time delay; the nonlinear receiving model is equivalent to a linear model based on the Bussgang decomposition principle; an expectation maximization algorithm framework is adopted, related parameters are jointly estimated through inner and outer layer iteration, an inner layer constructs a factor graph model based on a Markov chain to update an equivalent channel and a user state, and an outer layer adopts a local search algorithm to update a transmission delay estimation value. The method is used for a 1-bit low-precision receiver in an asynchronous large-scale machine type communication system, and realizes joint acquisition of active user detection, transmission delay estimation and channel state information.

The invention discloses a robot trajectory learning generation method and system based on an improved k value selection algorithm, and belongs to the technical field of robot control and trajectory planning. Comprising the steps of obtaining original trajectory data in a robot teaching process, and performing preprocessing; determining an optimal Gaussian kernel number by adopting an improved k value selection algorithm; executing k-means clustering by using the determined optimal k value to obtain an initial mean vector, a covariance matrix and a weight parameter of a Gaussian mixture model; an expectation maximization algorithm is adopted to carry out iterative calculation on the Gaussian mixture model, the iterative calculation obtains the posterior probability of each data point based on Gaussian distribution based on the step E, and the posterior probability is utilized to re-estimate the parameters of the Gaussian mixture model based on the step M; and according to specific task requirements, the expected planning trajectory of the robot is generated through Gaussian mixture regression by using the trained Gaussian mixture model parameters, so that the accuracy of robot trajectory generation is improved.

A photoelectron spectroscopy method includes: performing photoelectron spectrum detection on a target sample to obtain an initial photoelectron spectrum; selecting spectrum data points according to the initial photoelectron spectrum to obtain a photoelectron spectrum data point sequence; performing a first model parameter update on a preset initial photoelectron spectrum fitting model according to a preset expectation-maximization algorithm and the photoelectron spectrum data point sequence to obtain a first photoelectron spectrum fitting model; acquiring a target quantum effect constraint for the target sample; performing a second model parameter update on the first photoelectron spectrum fitting model according to a preset central field approximation relation and the target quantum effect constraint to obtain a second photoelectron spectrum fitting model; and performing spectrum fitting on the photoelectron spectrum data point sequence according to the second photoelectron spectrum fitting model to obtain a target photoelectron spectrum.

The invention relates to the technical field of mineral processing engineering, and discloses an ore grinding granularity prediction method combining missing value completion and multi-model collaboration. According to the method, a missing feature complementation module is constructed based on a Gaussian mixture model and an expectation maximization algorithm, and a linear distribution regression branch and a gradient lifting branch are constructed based on a linear regression model and a gradient lifting model respectively, so that an ore grinding granularity prediction model is formed. The method comprises the following steps: acquiring and preprocessing multi-source ore grinding operation data to obtain input data containing missing features, and carrying out probability modeling and completion on the missing features by utilizing a completion module; and the linear distribution regression branch and the gradient lifting branch perform modeling on the completion features respectively, output corresponding ore grinding particle size probability distribution prediction results, and fuse multi-branch prediction results to obtain final ore grinding particle size prediction distribution. According to the method, the robustness, precision and prediction stability of ore grinding particle size prediction under complex working conditions are effectively improved.

The application discloses a line intelligent planning and fault early warning method suitable for power engineering design and relates to the technical field of power engineering. The method comprises the following steps: through spatial semantic segmentation and a cost surface model, quantifying geographic information, geological conditions and construction cost into a continuous decision surface, combining a dynamic clustering algorithm to identify a large-area low-cost area as a primary feasible region; based on spatial features, geological stability and facility correlation data, constructing multi-dimensional decision indicators, and through real-time correlation coefficients, identifying "fault regions" such as terrain mutations or facility conflicts, and triggering an abnormal processing mechanism; selecting a seed region with the optimal construction condition as an anchor point, constructing a joint probability model of spatial features and cost, combining an expectation maximization algorithm and graph optimization technology, and generating an optimal path that takes into account the feasibility probability and cost benefit. The application improves the scientificity and economy of cable planning under complex geological conditions and provides key technical support for the intelligentization of power engineering.

The invention discloses a load modeling and regulating method and device participating in demand response, equipment and a medium. The method comprises the steps that firstly, operation parameters of a partial monitoring air conditioner are collected, and the single adjusting capacity is calculated based on a second-order equivalent thermal parameter model; quantifying cluster parameter heterogeneity by using a Gaussian mixture model, optimizing parameters through K-means and expectation maximization algorithms, and extrapolating the capacity to obtain a total adjustable capacity; then constructing a Markov chain state space containing a temperature interval, an action and a counter and a transfer matrix; establishing a function taking a power grid demand tracking error and economical efficiency as targets, and solving by adopting an alternating direction multiplier method to obtain the optimal temperature rise probability of each temperature interval; and finally generating a low-dimensional broadcast signal to guide the local autonomous decision of the air conditioner cluster. According to the method, the problems of accurate modeling and cooperative control of large-scale heterogeneous air conditioner loads are solved, accurate evaluation of the adjusting capacity and accurate guidance of group response are achieved, and communication and calculation expenses are remarkably reduced.

The invention discloses a parameter and state joint estimation method for inputting a nonlinear model under non-ideal data, and belongs to the technical field of industrial process control, system identification and state estimation. According to the method, a joint estimation algorithm based on an expectation maximization algorithm, unbiased finite impulse response filtering and a probability weighted delay mechanism is provided for the common concurrent defects of input nonlinearity, abnormal values, random time delay, data missing and the like in the industrial process. By introducing measurement noise containing abnormal values in student t distribution modeling and regarding missing data, time delay and the like as hidden variables, alternate updating of parameters and states is realized under an EM algorithm framework. The core of the method is that a probability weighted delay PWD mechanism is provided, soft weighting is carried out by using complete posteriori distribution of delay, and robustness to delay uncertainty is enhanced; and meanwhile, a UFIR filter which does not need priori noise statistical information is adopted to carry out state estimation, so that non-Gaussian noise interference is effectively resisted.

The invention provides a crude oil cutting calculation method, a crude oil cutting calculation system and a computer readable storage medium. The crude oil cutting calculation method comprises the following steps: acquiring an oil product type and oil product distillation data; based on the oil product type and the oil product distillation data, determining a plurality of Weibull distributions forming mixed Weibull distributions, the Weibull coefficients of the plurality of Weibull distributions being preset constants; calling a least square method to calculate the cumulative probability density function of the plurality of Weibull distributions so as to update the Weibull coefficients of the plurality of Weibull distributions; determining mixed Weibull distribution according to the Weibull coefficients of the plurality of Weibull distributions and the corresponding weighting coefficients; performing iterative adjustment on the Weber coefficients of the plurality of Weber distributions and the corresponding weighting coefficients through an expectation maximization algorithm to update the mixed Weber distribution; and determining the oil product distribution of the crude oil according to the updated mixed Weibull distribution.

A method for simultaneously deducing single-cell pseudo time, velocity field and gene interaction comprises the following steps: preprocessing a single-cell RNA sequencing data set of cell development with multiple branches, clustering different cell types, constructing a piecewise linear model, optimizing the piecewise linear model by adopting an expectation maximization algorithm, and calculating the single-cell pseudo time, velocity field and gene interaction. Deduced pseudo time of each cell, a gene interaction matrix of the dynamic network and single cell velocity field visualization on a pseudo time chart are obtained. According to the method, a segmented ordinary differential equation model is introduced to reconstruct an RNA velocity field and pseudo time of a cell and an interaction network between genes. By iteratively optimizing a connection matrix between genes and pseudo time of the cells, the prediction precision of dynamic change of the cells can be remarkably improved, cell state transition can be accurately deduced, and a key gene regulation and control relationship can be accurately detected.

This invention discloses an image line intersection detection method, system, and readable storage medium. The method includes: acquiring an input original image and performing multi-scale gradient field calculation to generate a weighted guided edge map; selecting seed points from the weighted guided edge map to generate a support region, and iteratively competing based on the support region to obtain two candidate point sets; constructing a dual-line-intersection joint probability model based on the two candidate point sets, wherein the points in the candidate point sets are used as latent variables to be estimated, and the point sets in the weighted guided edge map are used as observation data; and solving the dual-line-intersection joint probability model using the expectation-maximization algorithm to obtain the detection result, wherein the detection result includes at least the coordinates of the image intersection points. This invention can significantly improve the accuracy and robustness of image line intersection detection, effectively overcoming the problems of traditional methods such as excessive dependence on edge detection quality, sensitivity to parameter settings, and weak ability to handle line segment breaks.

The application discloses a bridge operation modal analysis method based on structural dynamics and Gaussian process fusion, and can be used for bridge structure health monitoring. Acceleration sensors are arranged at key positions of the bridge to collect vibration responses in an operation state; mode force power spectral densities of each mode are extracted through frequency domain analysis as external excitation strength indexes. The indexes are taken as Gaussian process inputs and hidden variables are introduced to quantize influences of excitation changes on natural frequencies, damping ratios and vibration modes; an expectation-maximization algorithm is used to iteratively identify structural dynamics parameters and Gaussian process hyperparameters. Finally, the bridge modal parameters after excitation disturbances such as wind and vehicle flow are removed and the hyperparameters representing working conditions are obtained, and long-term tracking, evaluation and prediction of the bridge structure health are realized.

The application provides a high-order interaction prediction method and device with mixed graph deep learning, which first constructs a drug molecule graph, a microorganism weighted graph, a disease weighted graph and a supergraph connecting the three based on multi-source heterogeneous data such as drug molecular structure, microorganism classification information and disease semantic network, forming a mixed graph structure. Subsequently, through a mixed graph deep learning module fusing a graph convolution network and a supergraph neural network, nonlinear structure features and high-order interaction features of each entity are extracted, and the adaptive fusion of the features is realized by using an attention mechanism. Then, the fused deep features are mapped to the prior expectation of the latent factor matrix in the Bayesian logic tensor decomposition model, a probabilistic graph model is constructed, and the joint adaptive inference of the model parameters, latent variables and deep learning mapping is carried out through a variational expectation maximization algorithm, so that the high-order correlation probability prediction of the whole tensor space is realized without negative sampling.

The invention discloses a satellite remote sensing image power tower disaster damage identification method and system and a related device. The method comprises the following steps: acquiring a multi-source power tower satellite remote sensing image and constructing a data set; inputting the data set into a pre-constructed rotating target detection model, outputting the spatial position of the electric power tower through the rotating target detection model, and identifying an orientation angle; according to the spatial position of the electric power tower and the identified orientation angle, constructing a maximum likelihood estimation model of angle distribution of the electric power tower through angle specialization processing, and iteratively solving parameters of the maximum likelihood estimation model by using an expectation maximization algorithm to determine orientation angle distribution of the electric power tower in the to-be-detected image; and performing anomaly detection according to the orientation angle distribution of the power tower in the to-be-detected image, and screening an object exceeding a threshold range, namely the power tower with disaster damage. The problem of sample scarcity can be avoided, the data preparation threshold is remarkably reduced, the detection process is fully automatic from data preprocessing to exception screening, and the disaster damage sensing efficiency and accuracy are effectively improved.

The invention provides an inertia and acoustics integrated navigation method based on enhanced hybrid minimum error entropy and unscented Kalman filtering. The method comprises the following steps: establishing a state equation model of an inertia and acoustics integrated navigation system; establishing an enhanced mixed minimum error entropy measurement model considering a sound ray bending effect and a multi-modal noise characteristic; based on a state equation model and an enhanced hybrid minimum error entropy measurement model, an EnMMEE-UKF framework is constructed, nonlinear mapping is processed through unscented transformation, an attenuation factor in a strong tracking filtering theory is introduced to correct an error covariance matrix, and an expectation maximization algorithm is adopted to adaptively adjust a hybrid coefficient of a hybrid kernel function. And state estimation and a covariance matrix are recursively updated in combination with a fixed point iteration method, so that real-time updating and feedback of errors are realized.

The present application provides a lightweight high-dimensional multivariate financial time series data joint prediction system and method thereof, relating to the technical field of time series data analysis, and solving the problems of complex neural network framework in the existing neural network, such as non-interpretability, complexity and large computational power consumption, including a data acquisition module, a data cleaning module, a model training module, a data prediction module and a prediction visualization unit. The method includes step 1, obtaining, by a data acquisition module, time series data from different financial sources. Compared with the traditional LDS model, the aLDS adopted by the present application has a novel parameter averaging process, and in the expectation maximization algorithm, the parameters obtained by training each financial time data are averaged, so as to achieve the purpose of training multiple financial time data together.

The application provides a position and posture capturing method and system applied to a motion capture system, and aims to solve the technical problem of low motion capture precision. The position and posture capturing method of the motion capture system comprises the following steps: obtaining a nominal position and posture transformation matrix and an actual position and posture transformation matrix through joint angles, end effector speed and end effector angular velocity; then, a position and posture deviation vector is calculated; the position and posture deviation vector is input into a least square matrix equation to obtain geometric error and random colored noise; based on an expectation maximization algorithm, the geometric error and the random colored noise are converged to obtain actual kinematic parameters; the actual kinematic parameter value is input into the motion capture system to obtain a calibrated motion capture system; the joint angles are input into the motion capture system to output a simulation posture and position, the error caused by the colored noise is reduced, and the calibration precision of the kinematic parameters is improved.

The application discloses a color size grouping and purchase template generation method based on Gaussian mixture clustering, and the method comprises the following steps: obtaining historical sales data; mapping all sizes into continuous size code values to form a one-dimensional size sample set; constructing a Gaussian mixture model containing multiple Gaussian distribution components; iteratively fitting the Gaussian mixture model through an expectation maximization algorithm to obtain a mixture weight, a mean value and a variance, and a posterior probability of each size sample belonging to each Gaussian distribution component; for each color identification, calculating the average posterior probability of the color identification belonging to each Gaussian distribution component, and distributing the color identification to the size structure group corresponding to the Gaussian distribution component with the maximum average posterior probability; based on all size samples in each group, counting the occurrence frequency of each size code value, calculating the proportion in the size structure group, and generating a size purchase template corresponding to the size structure group. The application realizes accurate purchase and inventory optimization.

The invention discloses a bridge operation modal analysis method based on structural dynamics and Gaussian process fusion, which can be used for bridge structure health monitoring. Arranging an acceleration sensor at a key position of the bridge, and collecting operation state vibration response; and extracting the modal force power spectral density of each modal through frequency domain analysis as an external excitation intensity index. Taking the index as Gaussian process input, introducing a hidden variable, and quantifying the influence of excitation change on the inherent frequency, the damping ratio and the vibration mode; and adopting an expectation-maximization algorithm to iteratively identify structural dynamic parameters and Gaussian process hyper-parameters. And finally, bridge modal parameters and hyper-parameters representing working conditions after excitation disturbance such as wind and traffic flow is eliminated are obtained, and long-term tracking, evaluation and prediction of bridge structure health are realized accordingly.

The invention discloses an oil reservoir interval water injection effect evaluation method and system based on PCA-GMM-XGBoost, and relates to the technical field of oilfield development, and the method comprises the steps: collecting and preprocessing multi-source development data of a target oil reservoir interval, forming an initial feature set, and extracting a key index system through principal component analysis; taking the key index system as input, initializing a Gaussian mixture model corresponding to the three-stage water injection effect, iteratively optimizing model parameters by using an expectation maximization algorithm, and completing quantitative classification of the water injection effect according to the posterior probability of each evaluation unit; taking the key index system as input, taking classification labels as targets, constructing and training an XGBoost multi-classification model containing regularization items, predicting and adjusting the direction of a new sample, and outputting a final decision after multi-factor rechecking is carried out on a fuzzy sample. According to the invention, by introducing advanced data analysis and machine learning technologies, accurate quantitative evaluation of the water injection effect and intelligent and scientific decision of the water injection adjustment strategy are realized.